Papers
6
Total Citations
150
H-Index
5
About
Sven Meier is a leading researcher in computer vision and robotics, whose work focuses on enabling machines to perceive and interact with the world at scale. His core research areas include category-level 6D object pose and shape estimation, robotic grasping, and self-supervised learning, with a particular emphasis on handling photometrically challenging objects like transparent or reflective items. Meier’s major contributions include pioneering deep learning approaches for class-level pose estimation, notably through his work on the CPS and CPS++ frameworks, which moved the field beyond instance-level methods to handle hundreds of object categories from monocular images. He has also created critical benchmark datasets, such as PhoCaL and HouseCat6D, which provide high-quality, multi-modal annotations for realistic scenarios, each garnering 26–45 citations. His work on DemoGrasp advances few-shot robotic grasping by leveraging human demonstrations, bridging the gap between perception and manipulation. With over 150 total citations, Meier’s research is shaping the future of scalable robotic perception, enabling robots to operate seamlessly in unstructured home environments through innovative self-supervised and language-guided learning techniques.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3DemoGrasp: Few-Shot Learning for Robotic Grasping with Human Demonstration29 citations · 2021
- 4
- 5CPS: Class-level 6D Pose and Shape Estimation From Monocular Images14 citations · 2020
- 6